an NPKtools White Paper

Unconscious Decisions in Market Pricing 1

Company Weighted vs. Incumbent Weighted Salary Distributions

By Lindsay Scott, Managing Partner, NPKtools

Question: Can you settle a dispute? One of our surveys provides both incumbent and company weighted market distributions. My colleague says we should use the incumbent distribution because it is a better representation of the “market”. I think the company weighted distribution is better. Who is right & what’s the difference between the two?

Answer: IT DEPENDS!!

Seriously though, if you’re market pricing jobs, the types of data you use can impact your pay recommendations, with implications for labor cost and competitiveness.

Just to ensure that we’re all on the same page, let’s define some terms:

Survey vendors that collect incumbent pay by job have the flexibility to create either type of market distribution. Survey vendors that collect only average pay by job produce company-weighted market distributions. While these vendors can provide an incumbent-weighted “average statistic,” it is extremely difficult for them to produce a true incumbent-weighted market distribution with accurate measures for all the market statistics (e.g. P25, P50, P75).

The answer to which type is best really depends on the three criteria below:

• Philosophy: who are you measuring your company against?
• Pay Strategy: what is your market comparison target, e.g. median (P50), P65, P25?
• Analytical Purpose: what question / issue will the pricing analysis address; annual planning, recruitment, retention or other?

Example Data

The best way to tackle this question is to create market distributions using both methodologies and plot the two distributions on top of each other. We’ll use real data from a large regional annual survey However, before we jump to that analysis let’s look at the raw data that makes up the market samples in the example. The graphic in Figure I summarizes both the range of employee/incumbent pay and company average pay for the seven firms in the survey.

Now let’s build the two distributions using both incumbent pay and company average pay, ranked from lowest to highest. Figure II shows the overlay for the two market distributions. The incumbent employee market distribution is represented by the line with blue diamonds. The company average pay market distribution is depicted by the colored circles. The incumbent distribution, of course, has a greater spread between lowest and highest pay values, but in this case company average pay is distributed nearly as widely as incumbent pay.

Figure II

The market statistics for the two distributions are summarized in Figure III. The company-weighted market (CWM) distribution has a median of $50,300. The ratio of P75 to P25 is 113.7%. The median for the incumbent-weighted market (IWM) distribution is lower by nearly $4,000. The IWM P75-25 ratio is higher at 120.7%. This is often but not always the case and will generally suggest that the IWM P75 will be higher than the P75 for CWM. The converse is also true; the IWM P25 will usually be lower than the CWM P25. In this case, except for the P75, the IWM statistics are all lower than the CWM. The difference ranges from $2,000 to $4,000 for this job market. IWM pay is lower here for two reasons:

• Company F has a high share (35.7%) of the total reported employees
• The pay practices for Company F, as reported in the survey, constrain employee pay to a fairly narrow band

Figure III

So, if the type of method used to compute market distributions can have such a significant impact, which should I use?

It Depends

To answer that question, let’s first summarize how the example distribution is calculated and then compare the two methodologies against the three criteria outlined above.

Company-Weighted Market Distribution

This distribution is calculated by averaging incumbent pay for employees within each company BEFORE constructing the market distribution. The distribution is based on the average pay for each of the seven firms. Using the average company values in the Figure I table, the median is the fourth observation (Company D-$50.3). The P75 is computed as the average of Company B & C. As can be seen comparing the table in Figure I with the company average positions in Figure II, the relative position for each firm has a substantial impact on the calculation for market statistics. As can also be seen in the Figure I table, the spread between the market endpoints (e.g. P25 and P75) is generally narrower than in the incumbent distribution. For this sample the ratio between P75 and P25 is 114%, versus 121% for the incumbent distribution.

Pay Philosophy: Use a company-weighted market distribution for comparing average pay for your company against the average company pay for other firms. In other words, your focus is on company to company competition. For this distribution, the P75 represents high paying companies whose average pay is higher than 75% of the firms in the sample. Because five of the firms are clustered together, the P50 & P75 are not substantially different.

Pay Strategy: Because the underlying data is averaged BEFORE the market distribution is computed, the spread between the P25 & P75 is narrower. If your pay strategy target is the P75 that means your firm’s labor costs will be lower if your employees on average are paid at parity to the P75 market. If you had 50 employees in this job, the fixed labor costs difference between the two methods would be $27,500. However, the reverse is also true. If your pay strategy is P25, then the company weighted distribution will generally yield higher market values.

Pricing purpose: This distribution is best for situations where you are comparing pay for all employees with a job rather than individuals or sub-groups. If senior management when executing a P75 strategy believe they are competing against high-paying firms rather than a 56 year old, career-level accountant with a high comparatio, then the company-weighted market data is more appropriate. Here, CWM serves as an indicator of competitive equity (“fairness”) since the company’s purpose in pricing generally is to maintain market equilibrium. Annual market planning for multiple incumbent jobs is a good example where the CWM methodology is probably more appropriate.

Incumbent-Weighted Market Distribution

This distribution is calculated by ranking all incumbent in the sample regardless of company affiliation and then calculating the market distribution

Pay Philosophy:This distribution measures your company pay against individuals in the market or more precisely company competition against people . In this distribution P75 represents high paid individuals with pay higher than 75% of the incumbents in the sample.

Pay Strategy:The distribution end-points (P10, P25, P75, P90) are generally higher or lower than the values of their company-weighted counterparts. If you are targeting the P75, the IWM distribution will generally produce a higher market level than a CWM distribution. In our sample case with 50 employees, the fixed labor cost impact is $27,500 higher. However for a P25 strategy the IWM in this case would lower fixed labor costs for 50 employees by $109,500.

Pricing purpose: The incumbent-weighted distribution is best in situations where you are competing in the market for individuals. Recruitment for individual or groups within hot skill markets is one example. In the case of recruitment, where you are trying to coax an incumbent to remove herself from the market, the incumbent-weighted distribution is a better measure of the possible range of pay that must be offered. In the case of dynamic or hot skill markets, individuals rather than companies create the market churn that raises market price levels. Since the churn impacts price levels at the margin, the extremities in the incumbent-weighted market distribution provide a better indicator of where prices at the margin may be heading and the price level that must be offered to retain your hot-market specialists. However, if markets are stable, then I would either use the CWM or compare the results of both distributions before making a recommendation.

So are we back to square 1?

Not really. While there is no single correct answer that is good for all situations, there are now three criteria that can help you determine which method is most appropriate for your current situation.